Researchers at the University of Edinburgh have developed a mathematical framework that could cut the energy required to store and manipulate digital information by several orders of magnitude. The method, published September 6 in Advanced Materials, uses optimal control theory to design magnetic-field pulses that flip bits with far less power than current memory technologies demand.
The work arrives as energy consumption from data centers continues to climb. AI models, recommendation engines, and large-scale scientific simulations all depend on vast amounts of data being moved, stored, and processed. Without efficiency gains, information and communication technologies could account for a substantial share of global electricity use in the coming decades.
Redesigning the pulse
Magnetic memory stores bits by switching magnetic states. The Edinburgh team approached this switching process as an optimization problem, applying a branch of mathematics called optimal control theory to find the most energy-efficient path between states.
The result is a framework for designing ultrafast magnetic-field pulses that minimize energy consumption while respecting real-world experimental constraints. Computer simulations suggest the technique could reduce switching energy far below what DRAM, STT-MRAM, and emerging SOT-MRAM devices currently require.
Closing in on a physical limit
The predicted energy requirements push future magnetic memory close to the Landauer limit - the fundamental thermodynamic minimum for processing a single bit of information. Approaching that boundary would represent a major step toward making computation as energy-efficient as physics allows.
"Every digital operation has an energy cost, and that cost becomes increasingly important as AI and data-intensive technologies continue to expand," said Dr. Elton Santos from the Institute for Condensed Matter Physics and Complex Systems, who led the research. "Our work shows that, by carefully designing how a magnetic field changes in time, magnetization can be switched far more efficiently than with conventional approaches."
The framework is not limited to magnetic fields. Santos said the same mathematics can be adapted to electrical currents and ultrafast laser pulses, two technologies under active investigation for next-generation data storage.
From theory to device
The paper goes beyond calculations. It includes guidance for experimental implementation, such as optimized device geometries and methods for delivering magnetic fields. These details could help other researchers test the concept in the lab.
The research builds on growing interest in van der Waals magnets, a class of materials that can be exfoliated into atomically thin layers. Their properties make them candidates for compact, low-power memory devices if switching can be controlled efficiently.
Why this matters for science and research professionals
For researchers working at the intersection of materials science, condensed matter physics, and computing hardware, this framework offers a new design principle: treat bit-flipping as an optimization problem rather than a brute-force operation. The open question is whether experimental groups can translate the simulations into functioning devices. If they can, the same optimal-control approach might extend to other areas where energy dissipation currently limits performance, from AI for Science & Research workloads to cryogenic computing. The mathematics, as Santos noted, is "far more versatile" than the magnetic-field case studied here.
Your membership also unlocks:
